[SYCL] Forward AOT backend/linker options as individual tokens by bader · Pull Request #22645 · intel/llvm

bader · 2026-07-15T17:18:15Z

Route -Xsycl-target-backend and -Xsycl-target-linker options through
--device-compiler=/--device-linker= uniformly for JIT and AOT SYCL
targets, one flag per token instead of joining tokens into a single
string. clang-linker-wrapper now keeps CLI-supplied AOT options as a
token list end-to-end and appends them to ocloc/opencl-aot argv
individually, avoiding a lossy re-split of values with embedded
spaces. JIT targets keep the existing string-joining behavior for the
image's compile-opts/link-opts properties.

Route -Xsycl-target-backend and -Xsycl-target-linker options through --device-compiler=/--device-linker= uniformly for JIT and AOT SYCL targets, one flag per token instead of joining tokens into a single string. clang-linker-wrapper now keeps CLI-supplied AOT options as a token list end-to-end and appends them to ocloc/opencl-aot argv individually, avoiding a lossy re-split of values with embedded spaces. JIT targets keep the existing string-joining behavior for the image's compile-opts/link-opts properties.

Copilot

Pull request overview

This PR adjusts SYCL option forwarding so that -Xsycl-target-backend / -Xsycl-target-linker are forwarded to clang-linker-wrapper as one --device-compiler= / --device-linker= flag per token (instead of joining tokens into a single space-separated string), and updates clang-linker-wrapper to preserve and forward those tokens end-to-end for AOT backends.

Changes:

Emit one --device-compiler=...=<token> / --device-linker=...=<token> per translated SYCL backend/link option in the driver.
Keep CLI-supplied AOT options as a token list in clang-linker-wrapper and append them directly to ocloc / opencl-aot argv (avoiding lossy re-splitting).
Update driver/wrapper tests to reflect the new per-token forwarding behavior (including AOT coverage).

Reviewed changes

Copilot reviewed 5 out of 5 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
clang/tools/clang-linker-wrapper/ClangLinkerWrapper.cpp	Thread AOT device args as token lists through the SYCL pipeline and append to AOT tool invocations.
clang/lib/Driver/ToolChains/Clang.cpp	Forward SYCL backend/linker options to `clang-linker-wrapper` as one flag per token.
clang/test/Driver/sycl-offload.cpp	Update checks to expect per-token `--device-compiler/--device-linker` forwarding.
clang/test/Driver/sycl-offload-new-driver.cpp	Add coverage ensuring AOT targets also forward backend/linker options as separate tokens.
clang/test/Driver/clang-linker-wrapper.cpp	Extend wrapper tests for split `-device` reconstruction and multi-arg AOT option forwarding.

+  std::string AOTOptions;
+  if (Triple.isSPIRAOT()) {
+    AOTOptions = CompileLinkOptions.first;
+    if (!CompileLinkOptions.second.empty()) {
+      if (!AOTOptions.empty())
+        AOTOptions += ' ';
+      AOTOptions += CompileLinkOptions.second;
+    }
+  }


Update the CHK-NO-CMDS-AOT-GEN-LINKERARG test to assert on the actual ocloc invocation instead of an unrelated sycl-post-link line, matching the new one-CLI-token-equals-one-argv-entry contract for --device-compiler=/--device-linker=. Remove CHK-NO-CMDS-AOT-GEN (relied on a single-token "-device pvc" CLI value that is no longer split, and is unreachable via the real driver since -Xsycl-target-backend already tokenizes it into separate flags) and CHK-MULTI-DEVICE-LINKER-AOT (duplicated the same reconstruction behavior as the fixed test).

-Xsycl-target-backend/-Xsycl-target-linker require the driver and clang-linker-wrapper to parse and re-serialize ocloc's own option syntax (e.g. the literal "-options " split in addOclocOptions), which is fragile — appended link-opts can be silently absorbed into the -options value. Note the deprecation/removal of these options as a follow-up to avoid this class of parsing entirely.

Add a comment cross-referencing Driver::getOffloadArchs, which performs a similar "-device" token detection on the driver side for spir64_gen Arch selection, so a future edit to either detector's handling of -Xsycl-target-backend syntax prompts a check of the other.

…Module Triple parsing, AOTOptions/BackendOptions derivation, and the -device detection in IsDevicePassedWithSyclTargetBackend were recomputed on every call to postLinkProcessModule, even though the value is identical across all split modules sharing a triple. Moved the computation up to runSYCLOffloadingPipeline, which now computes it once before looping over Modules and passes the results down.

invokeBackendForSYCLDevice and compileDeviceAndBundle are single-caller helpers whose only call sites already pass AOTDeviceArgs explicitly, so the `= {}` defaults were unreachable and misleadingly implied some caller omits the argument.

BuildArgs is joined into a single space-separated compile-opts=/link-opts= string before being embedded in the offload image, then later re-split on spaces by clang-linker-wrapper for AOT triples. This corrupts any token containing an embedded space unless it happens to fall inside the pre-existing "-options \"...\"" wrapper convention. The CLI-supplied counterpart (--device-compiler=/--device-linker=) already avoids this by forwarding individual tokens via AOTDeviceArgs.

linkAndWrapDeviceFiles() and runSYCLOffloadingPipeline() each decide independently whether CLI-supplied AOT options go through the token vector or the flat-string path, with nothing enforcing the two isSPIRAOT() checks stay in agreement.

bader · 2026-07-17T17:23:46Z

ping

mdtoguchi · 2026-07-17T18:47:06Z

    CmdArgs.push_back("-options");
    CmdArgs.push_back(AfterOptions);
  }
+  llvm::append_range(CmdArgs, AOTDeviceArgs);


Is there any concern on ordering of the options? AOTDeviceArgs will always fall after the other options.

I'm not aware of any issues with ocloc argument ordering.
Also, I don't think my patch changes the ordering.

llvm/clang/tools/clang-linker-wrapper/ClangLinkerWrapper.cpp

Lines 2459 to 2481 in b9d460f

Expected<std::pair<std::string, std::string>> CompileLinkOptionsOrErr =

extractSYCLCompileLinkOptions(Input);

if (!CompileLinkOptionsOrErr)

return CompileLinkOptionsOrErr.takeError();

std::pair<std::string, std::string> &CompileLinkOptions =

*CompileLinkOptionsOrErr;

// Append device compiler and linker options passed via

// -device-compiler= and -device-linker= to clang-linker-warpper,

// together with options extracted from the image.

StringRef DeviceCompilerArgs =

LinkerArgs.getLastArgValue(OPT_compiler_arg_EQ);

if (!DeviceCompilerArgs.empty()) {

CompileLinkOptions.first += " ";

CompileLinkOptions.first += DeviceCompilerArgs;

}

StringRef DeviceLinkerArgs =

LinkerArgs.getLastArgValue(OPT_linker_arg_EQ);

if (!DeviceLinkerArgs.empty()) {

CompileLinkOptions.second += " ";

CompileLinkOptions.second += DeviceLinkerArgs;

}

Do you think there is a problem with the order of ocloc arguments?

I suppose not - dividing up the backend args and device args shouldn't provide opportunity for potential override behaviors

mdtoguchi

OK by me - lots of TODO and FIXME items that need to be tracked.

sys-ce-bb · 2026-07-18T00:31:40Z

@intel/llvm-gatekeepers please consider merging

YuriPlyakhin · 2026-07-18T02:36:04Z

-// (happen when AOT device is specified via -Xsycl-target-backend '-device pvc' in clang), 
-// the target is not passed to sycl-post-link for filtering.
-// RUN: clang-linker-wrapper -sycl-embed-ir --bitcode-library=spir64_gen-unknown-unknown=%t1.devicelib.bc -sycl-post-link-options="SYCL_POST_LINK_OPTIONS" -llvm-spirv-options="LLVM_SPIRV_OPTIONS" "--host-triple=x86_64-unknown-linux-gnu" "--device-compiler=spir64_gen-unknown-unknown=-device pvc" "--linker-path=/usr/bin/ld" "--" HOST_LINKER_FLAGS "-dynamic-linker" HOST_DYN_LIB "-o" "a.out" HOST_LIB_PATH HOST_STAT_LIB %t1.o --dry-run 2>&1 | FileCheck -check-prefix=CHK-NO-CMDS-AOT-GEN %s
-// CHK-NO-CMDS-AOT-GEN: sycl-post-link{{.*}} SYCL_POST_LINK_OPTIONS -o {{[^,]*}}.table {{.*}}.bc


The removed CHK-NO-CMDS-AOT-GEN check verified that -device suppresses the intel_gpu_... prefix passed to sycl-post-link. The IsDevicePassedWithSyclTargetBackend path is now untested.
Please add a sycl-post-link{{.*}} -o {{[^,]*}}.table check to the new RUN line.

YuriPlyakhin · 2026-07-18T02:42:45Z

-// Check that in AOT case backend options are passed to ocloc and are not passed to offload wrapper
-// because SYCL Runtime can't make any use of it in AOT case.
-// CHECK-COMPILE-LINK-OPTS-AOT: ocloc{{.*}} -device pvc ccc ccc -output
+// Check that in AOT case backend and linker options are passed to ocloc and


This PR also changes AOT behavior: image-embedded link-opts are now forwarded to ocloc, whereas before only compile-opts reached the AOT backend. This is reasonable, but it's a behavior change independent of the token-splitting fix — please call it out in the PR description.

YuriPlyakhin · 2026-07-18T02:49:56Z

+// Check that when "-device pvc" specification, split across two separate
+// --device-linker= arguments (each --device-compiler=/--device-linker= CLI
+// occurrence becomes exactly one ocloc argv entry, so a multi-token value
+// like "-device pvc" must be supplied as separate occurrences, not as one
+// value containing an embedded space), is reconstructed correctly in the
+// ocloc invocation.


nit

Suggested change

// Check that when "-device pvc" specification, split across two separate

// --device-linker= arguments (each --device-compiler=/--device-linker= CLI

// occurrence becomes exactly one ocloc argv entry, so a multi-token value

// like "-device pvc" must be supplied as separate occurrences, not as one

// value containing an embedded space), is reconstructed correctly in the

// ocloc invocation.

// Check that a "-device pvc" specification split across two separate

// --device-linker= arguments (each --device-compiler=/--device-linker= CLI

// occurrence becomes exactly one ocloc argv entry, so a multi-token value

// must be supplied as separate occurrences)

// is reconstructed correctly in the ocloc invocation.

bader added 2 commits July 15, 2026 10:14

clang-format

131a671

bader requested a review from Copilot July 15, 2026 17:37

Copilot started reviewing on behalf of bader July 15, 2026 17:37 View session

Copilot AI reviewed Jul 15, 2026

View reviewed changes

bader added 7 commits July 15, 2026 15:46

bader marked this pull request as ready for review July 16, 2026 00:26

bader requested review from a team as code owners July 16, 2026 00:26

mdtoguchi reviewed Jul 17, 2026

View reviewed changes

mdtoguchi mentioned this pull request Jul 17, 2026

[SYCL][NewOffloadModel] Fix -Xdevice-post-link, -Xspirv-translator and -Xspirv-to-ir-wrapper #22664

Open

mdtoguchi approved these changes Jul 17, 2026

View reviewed changes

uditagarwal97 merged commit dd9abc1 into intel:sycl Jul 18, 2026
31 checks passed

YuriPlyakhin reviewed Jul 18, 2026

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

[SYCL] Forward AOT backend/linker options as individual tokens#22645

[SYCL] Forward AOT backend/linker options as individual tokens#22645
uditagarwal97 merged 9 commits into
intel:syclfrom
bader:sycl-aot-token-fix

bader commented Jul 15, 2026

Uh oh!

Copilot AI left a comment

Uh oh!

bader commented Jul 17, 2026

Uh oh!

mdtoguchi Jul 17, 2026

Uh oh!

bader Jul 17, 2026

Uh oh!

mdtoguchi Jul 17, 2026

Uh oh!

mdtoguchi left a comment

Uh oh!

sys-ce-bb commented Jul 18, 2026

Uh oh!

Uh oh!

YuriPlyakhin Jul 18, 2026 •

edited

Loading

Uh oh!

YuriPlyakhin Jul 18, 2026

Uh oh!

YuriPlyakhin Jul 18, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

6 participants

	Expected<std::pair<std::string, std::string>> CompileLinkOptionsOrErr =
	extractSYCLCompileLinkOptions(Input);
	if (!CompileLinkOptionsOrErr)
	return CompileLinkOptionsOrErr.takeError();

	std::pair<std::string, std::string> &CompileLinkOptions =
	*CompileLinkOptionsOrErr;

	// Append device compiler and linker options passed via
	// -device-compiler= and -device-linker= to clang-linker-warpper,
	// together with options extracted from the image.
	StringRef DeviceCompilerArgs =
	LinkerArgs.getLastArgValue(OPT_compiler_arg_EQ);
	if (!DeviceCompilerArgs.empty()) {
	CompileLinkOptions.first += " ";
	CompileLinkOptions.first += DeviceCompilerArgs;
	}
	StringRef DeviceLinkerArgs =
	LinkerArgs.getLastArgValue(OPT_linker_arg_EQ);
	if (!DeviceLinkerArgs.empty()) {
	CompileLinkOptions.second += " ";
	CompileLinkOptions.second += DeviceLinkerArgs;
	}

Uh oh!

Conversation

bader commented Jul 15, 2026

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

bader commented Jul 17, 2026

Uh oh!

mdtoguchi Jul 17, 2026

Choose a reason for hiding this comment

Uh oh!

bader Jul 17, 2026

Choose a reason for hiding this comment

Uh oh!

mdtoguchi Jul 17, 2026

Choose a reason for hiding this comment

Uh oh!

mdtoguchi left a comment

Choose a reason for hiding this comment

Uh oh!

sys-ce-bb commented Jul 18, 2026

Uh oh!

Uh oh!

YuriPlyakhin Jul 18, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

YuriPlyakhin Jul 18, 2026

Choose a reason for hiding this comment

Uh oh!

YuriPlyakhin Jul 18, 2026

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

6 participants

YuriPlyakhin Jul 18, 2026 •

edited

Loading